Laut MIT-Studie kann KI bereits 11,7 % der US-Arbeitskräfte ersetzen

The part they left out is the 11.7 % are all CEOs

Can someone enter on https://iceberg.mit.edu/? I can’t, and I’m curious how they can tell a „skill“ can be done with AI.

I’d like to see some clear data and case studies where AI HAS replaced workers. The previous MIT study (ie 95% of AI deployments fail) cited actual data. Whereas this work appears to be a model based on assumptions as to why AI can readily accomplish. The fundamental issue with that is there aren’t many examples of AI accomplishing those tasks.

Show statistically significant data demonstrating what is claimed and I’ll believe it. Otherwise, these projections are as worthless as any of the other AI ones.

I don’t understand?
Blue collar jobs have been Automated for the last 100 years, from containers at the docks, powertools , then computers
How many secretary pools are still here, 1000sof jobs lost to computers.
Mobile phones, speak straight to person not leave msg.

I would ask these researchers to build and productionalize and agent. Let’s see how this changes opinions. CEO’s are jumping on this bandwagon without taking into consideration the training of the next generation. Yes it is ‘saving’ money, but is it really. They also don’t share how extremely difficult it is to build prod ready AI.

Spoiler: Rage at shouting at an AI Chatbot that won’t let you speak to a human will increase 111.7%.

While AI can replace 11.7% of the workforce, it will do it poorly and lower customer satisfaction. Not that CEO’s care as long as the shareholders are happy.

just because you CAN do something doesn’t mean you should. AI won’t earn a wage and put that money back into the economy.

And the mass layoffs continue

And the stock market rallies

And the CS major can’t find a job so they go back to get a CCNP and work as a Sys…ooops that job just got canned by management as Co-pilot can now monitor EC2.

Fucking co-pilot.

This study comes from the MIT Media Lab, which has a long history of funding from illustrious philanthropists such as noted pedophile Jeffrey Epstein and de facto leader of Saudi Arabia Mohammed bin Salman Al Saud.

Makes you wonder who exactly is financing studies like this and what heinous crimes they will be accused of in the future!

Funded by MIT grads in Palo Alto

I look forwards to all the new jobs that will be created to clean up shit made by AI.

I’m already using it in my every day workflow. It’s only a matter of time until it learns how to do my job without me. Many entry level and mid management jobs are going away and have already begun to

“Can”
But where has it and what has been the result?

I feel like on the base level it can replace many roles but only for portions of their work. Not fully.

Aren’t we always trying to create more jobs? What’s this fascination with creating fewer jobs or replacing jobs?

Flash news: AI can take over AI jobs.

bull dookie. ai ain’t shit compared to just plain old i.

I thought jobs were good for our economy?

No, firing people is good for the economy I guess.

Another way to phrase it – MIT study finds 11.7% of US workforce doesn’t actually do anything all day.

It’s probably more like double that if we’re being honest. A lot of office work is straight-up bullshit that exists just so somebody can fill a seat 8 hours a day.

Did they use AI to determine that number?

No it cannot.

AI can be forced into taking positions, but AI (LLMs) are not *good* at handling an entire job. Most jobs require communication with other individuals and also processing new information we may receive.

AI is extremely bad at understanding new information and often hallucinates because it cannot understand a “truth”, it only does what LLMs do.

It can’t replace taking orders at Taco Bell. It can’t handle call trees often because its ability to process language is worse than the average individual. It can’t provide information which is an iron clad truth. LLM’s are exceedingly expensive that are heavily subsidized by businesses pouring money into it. That money runs out someday.

Recently, on Ed Zitron’s podcast, he stated that OpenAI would need to have as many subscribers as Netflix (the biggest streaming service in the world) each paying something like $200/mo to make the service break in on costs right now.

The cost of AI goes up dramatically by the month. Electricity, memory chips, GPU’s.

This isn’t sustainable.

BREAKING:

Results of nerd project at insular nerd school shows 11.7% of non nerd jobs can be replaced by nerd invention. Critics of the claim state that the nerd invention doesn’t work very well, and only seems capable of moving money from the pockets of investors into a large gaping hole in the economy where hopes and dreams go to die. The nerds fired back, stating that the invention doesn’t need to work very well to replace non-nerd jobs, because nerds are the only ones who do real work anyways.

„We need to get in front of this asap,“ says one giant nerd attached to the project. „That 11.7% is comprised almost entirely of jocks, bullies, preps, chads, etc. Without make work jobs their attention will turn elsewhere.“

„I of course refer to the sensitive intellectual, such as myself.“ the nerd continued. „If we don’t act now, there won’t be enough toilets in the continental US to support the amount of swirlies that will be occurring in 18 months time.“

„It’s clear that posting about it on Less Wrong is no longer sufficient. We must invest in a solution to this problem.“

The nerd then went on to describe their solution, which was basically just Nazism with a racing stripe painted on it.